Public perspectives on consent for the linkage of data to evaluate vaccine safety

We sought community opinion on consent alternatives when linking childhood immunisation and hospital attendance records for the purpose of vaccine safety surveillance.

We conducted computer-assisted telephone interviewing (CATI) of a sample of rural and metropolitan residents of South Australia in 2011.

Of 2002 households interviewed (response rate 55.6%), 96.4% supported data linkage for postmarketing surveillance of vaccines; very few were completely opposed (1.5%) or undecided (2.2%). The majority (75.3%) trusted the privacy protections used in data linkage and most wished to have minimal or no direct involvement, preferring either opt-out consent (40.4%) or no consent (30.6%). A quarter of respondents (24.6%) favoured opt-in consent, but their preferences were divergent; half requested consent be sought prior to every use (11.4%) while the remainder preferred to give broad consent just once (3.4%) or renewed at periodic intervals (9.8%). Over half of the respondents gave higher priority to rapid vaccine safety surveillance (56.5%) rather than first seeking parental consent (26.6%) and one in seven was undecided (14.5%). Although 91.6% of respondents believed childhood vaccines are safe, over half (53.1%) were very or somewhat concerned that a vaccine could cause a serious reaction. Nevertheless, 92.4% of the parents in the sample (556/601) reported every child in their care as being fully immunised according to the National Immunisation Program schedule. Only 3.7% of parents (22/601) reported one or more children as under immunised, and 3.9% (23/601) reported that none of their children were immunised.

This survey demonstrates that data linkage for vaccine safety surveillance has substantial community support and that a system utilising opt-out consent or no consent was preferred to one using opt-in consent. These findings should inform public health policy and practice; data linkage should be established where feasible to address limitations in passive surveillance systems.